3D surface pressure measurement with single light-field camera and pressure-sensitive paint

Abstract

A novel technique that simultaneously measures three-dimensional model geometry, as well as surface pressure distribution, with single camera is demonstrated in this study. The technique takes the advantage of light-field photography which can capture three-dimensional information with single light-field camera, and combines it with the intensity-based pressure-sensitive paint method. The proposed single camera light-field three-dimensional pressure measurement technique (LF-3DPSP) utilises a similar hardware setup to the traditional two-dimensional pressure measurement technique, with exception that the wind-on, wind-off and model geometry images are captured via an in-house-constructed light-field camera. The proposed LF-3DPSP technique was validated with a Mach 5 flared cone model test. Results show that the technique is capable of measuring three-dimensional geometry with high accuracy for relatively large curvature models, and the pressure results compare well with the Schlieren tests, analytical calculations, and numerical simulations.

Fast and Accurate 3D Measurement Based on Light-Field Camera and Deep Learning

“…The camera has a pixel resolution of 8000 × 5500 and MLA resolution of 1200 × 900. The camera is calibrated following the procedures introduced in [26]. In the measurement process, the system first captures raw light-field images of the measured object, which are the collection of individual images beneath each micro-lens of the whole MLA, and then generates multi-perspective SAIs from raw light-field images by taking the same pixel beneath each micro-lens and combining them in the same order as that of the micro-lens they are from [5].…”

“…Based on these advances, Ng et al [1] proposed a micro-lens array (MLA) based light-field camera where a high-resolution MLA was installed between the main lens and the camera sensor to record the plenoptic function of incoming light rays. Taking advantage of such light-field cameras, researchers have proven the feasibility of applying single light-field camera for 3D object reconstruction tasks in volumetric flow measurement and 3D geometry reconstruction [23,24,25,26].…”

Light-field imaging for distinguishing fake pedestrians using convolutional neural networks

“…23 Shi et al introduced a novel technique that can simultaneously measure 3D model geometry and 3D surface pressure with a single LF camera. 24 Wang et al proposed several novel CNN architectures to recognize the material on a 4D LF material dataset. 25 Fan and Yang estimated the depth of real-world scenes containing an object semisubmerged in water using a LF camera.…”

Detecting Small Scale Pedestrians and Anthropomorphic Negative Samples Based on Light-Field Imaging

“…Recent years, many new light-field camera applications have been put forward. Shi et al [24] introduced a novel technique that could simultaneously measure 3D model geometry and 3D surface pressure with a single light-field camera. Zhao et al [25] believed that light-field camera could be applied to depth extraction of low-texture region.…”